Analyzing a blast result of NGS data

Question

I am analyzing BLAST results from total RNA NGS, pair ends. This BLAST results was performed against the ssRNA virus database download from NCBI.

I filtered those read data whose paired ends uniquely matched a single virus and I have both (removing single reads). Moreover, based on the sstart and send values from a unique virus in each library I calculated the difference (ssdif) of the lowest (sstart) of the highest (send) value that I got in the BLAST results,

for example: If I got 4 reads that match against one virus in one library (2 reads with pair ends) I got the lowest sstart and the highest send for these four hits.

Then, I calculated the coverage using the ssdif; coverage=(ssdif/genome length for that virus) * 100

After I calculate the coverage for each virus in each library based of the BLAST results I calculated the mean coverage for each genus and for each family of all viruses detected in my analysis.

Questions

1. Does it make sense the way that I calculated the coverage? I would like to filter my data for those viruses that have similar or higher coverage to the genus or family (previously calculated)
   2. Based on what taxonomic level does it make more sense filter them?

Thanks in advance